Accuracy of a deep convolutional neural network in the detection of myopic macular diseases using swept-source optical coherence tomography

This study examined and compared outcomes of deep learning (DL) in identifying swept-source optical coherence tomography (OCT) images without myopic macular lesions [i.e., no high myopia (nHM) vs. high myopia (HM)], and OCT images with myopic macular lesions [e.g., myopic choroidal neovascularization (mCNV) and retinoschisis (RS)]. A total of 796 SS-OCT images were included in the study as follows and analyzed by k-fold cross-validation (k = 5) using DL’s renowned model, Visual Geometry Group-16: nHM, 107 images; HM, 456 images; mCNV, 122 images; and RS, 111 images (n = 796). The binary classification of OCT images with or without myopic macular lesions; the binary classification of HM images and images with myopic macular lesions (i.e., mCNV and RS images); and the ternary classification of HM, mCNV, and RS images were examined. Additionally, sensitivity, specificity, and the area under the curve (AUC) for the binary classifications as well as the correct answer rate for ternary classification were examined. The classification results of OCT images with or without myopic macular lesions were as follows: AUC, 0.983; sensitivity, 0.953; specificity, 0.940. The classification results of HM images and images with myopic macular lesions were as follows: AUC, 0.976; sensitivity, 0.940; specificity, 0.941. The correct answer rate in the ternary classification of HM images, mCNV images, and RS images were as follows: HM images, 93.7%; mCNV images, 82.4%; and RS, 92.3% with mean, 91.4%. Using noninvasive, easy-to-obtain swept-source OCT images, the DL model was able to classify OCT images without myopic macular lesions and OCT images with myopic macular lesions such as mCNV and RS with high accuracy. The study results suggest the possibility of conducting highly accurate screening of ocular diseases using artificial intelligence, which may improve the prevention of blindness and reduce workloads for ophthalmologists.

[1]  H. Taylor,et al.  Prevalence of Visual Impairment in the Adult Japanese Population by Cause and Severity and Future Projections , 2010, Ophthalmic epidemiology.

[2]  Lazaros S. Iliadis,et al.  Artificial Neural Networks - ICANN 2010 - 20th International Conference, Thessaloniki, Greece, September 15-18, 2010, Proceedings, Part I , 2010, International Conference on Artificial Neural Networks.

[3]  K. Naidoo,et al.  Global Prevalence of Myopia and High Myopia and Temporal Trends from 2000 through 2050. , 2016, Ophthalmology.

[4]  F. Wen,et al.  Fluorescein Leakage within Recent Subretinal Hemorrhage in Pathologic Myopia: Suggestive of CNV? , 2018, Journal of ophthalmology.

[5]  Sven Behnke,et al.  Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition , 2010, ICANN.

[6]  K. Nishida,et al.  Postoperative optical coherence tomographic appearance and relation to visual acuity after vitrectomy for myopic foveoschisis. , 2013, American journal of ophthalmology.

[7]  Thomas B. Moeslund,et al.  Long-Term Occupancy Analysis Using Graph-Based Optimisation in Thermal Imagery , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[8]  E. Topol,et al.  A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis. , 2019, The Lancet. Digital health.

[9]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[10]  T. Ninomiya,et al.  Original articlePrevalence and Risk Factors for Myopic Retinopathy in a Japanese Population: The Hisayama Study , 2012 .

[11]  A. Iwase,et al.  Prevalence and causes of low vision and blindness in a Japanese adult population: the Tajimi Study. , 2006, Ophthalmology.

[12]  Geraint Rees,et al.  Clinically applicable deep learning for diagnosis and referral in retinal disease , 2018, Nature Medicine.

[13]  Ya-Hsin Kung,et al.  FIVE-YEAR OUTCOMES OF INTRAVITREAL INJECTION OF RANIBIZUMAB FOR THE TREATMENT OF MYOPIC CHOROIDAL NEOVASCULARIZATION , 2017, Retina.

[14]  Gretchen A. Stevens,et al.  Causes of vision loss worldwide, 1990-2010: a systematic analysis. , 2013, The Lancet. Global health.

[15]  Richard F Spaide,et al.  Optical coherence tomography: imaging of the choroid and beyond. , 2013, Survey of ophthalmology.

[16]  David J. Fleet,et al.  Computer Vision – ECCV 2014 , 2014, Lecture Notes in Computer Science.

[17]  S. Fiske,et al.  The Handbook of Social Psychology , 1935 .

[18]  T. Toyo’oka,et al.  Automatic analyzer for highly polar carboxylic acids based on fluorescence derivatization-liquid chromatography. , 2015, Biomedical chromatography : BMC.

[19]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[20]  F. Wen,et al.  Long-Term Natural Outcomes of Simple Hemorrhage Associated with Lacquer Crack in High Myopia: A Risk Factor for Myopic CNV? , 2018, Journal of ophthalmology.

[21]  Y. Ikuno,et al.  Diagnosis and treatment guideline for myopic choroidal neovascularization due to pathologic myopia , 2017, Progress in Retinal and Eye Research.

[22]  H. Tabuchi,et al.  Automated detection of a nonperfusion area caused by retinal vein occlusion in optical coherence tomography angiography images using deep learning , 2019, PloS one.

[23]  J. Hanley,et al.  The meaning and use of the area under a receiver operating characteristic (ROC) curve. , 1982, Radiology.

[24]  K. Ohno-Matsui,et al.  FIVE-YEAR OUTCOMES OF INTRAVITREAL RANIBIZUMAB FOR CHOROIDAL NEOVASCULARIZATION IN PATIENTS WITH PATHOLOGIC MYOPIA. , 2017, Retina.

[25]  A. Erginay,et al.  Long-term follow-up of high myopic foveoschisis: natural course and surgical outcome. , 2007, American journal of ophthalmology.

[26]  Yoshinori Mitamura,et al.  Deep Neural Network-Based Method for Detecting Central Retinal Vein Occlusion Using Ultrawide-Field Fundus Ophthalmoscopy , 2018, Journal of ophthalmology.

[27]  T. Lai,et al.  LONG-TERM OUTCOMES OF RANIBIZUMAB TREATMENT OF MYOPIC CHOROIDAL NEOVASCULARIZATION IN EAST-ASIAN PATIENTS FROM THE RADIANCE STUDY , 2017, Retina.

[28]  Jost B Jonas,et al.  Prevalence and progression of myopic retinopathy in Chinese adults: the Beijing Eye Study. , 2010, Ophthalmology.

[29]  Faraz Oloumi,et al.  Deep learning in ophthalmology: a review. , 2018, Canadian journal of ophthalmology. Journal canadien d'ophtalmologie.

[30]  Yoshinori Mitamura,et al.  Accuracy of ultrawide-field fundus ophthalmoscopy-assisted deep learning for detecting treatment-naïve proliferative diabetic retinopathy , 2019, International Ophthalmology.

[31]  Gezhi Xu,et al.  RAPID PROGRESSION OF FOVEOMACULAR RETINOSCHISIS IN YOUNG MYOPICS. , 2019, Retina.

[32]  Y. Tano,et al.  Intravitreal bevacizumab for choroidal neovascularization attributable to pathological myopia: one-year results. , 2009, American journal of ophthalmology.

[33]  T. Yatagai,et al.  In vivo high-contrast imaging of deep posterior eye by 1-microm swept source optical coherence tomography and scattering optical coherence angiography. , 2007, Optics express.

[34]  Hideharu Ohsugi,et al.  Comparison between support vector machine and deep learning, machine-learning technologies for detecting epiretinal membrane using 3D-OCT , 2018, International Ophthalmology.

[35]  Hideharu Ohsugi,et al.  Accuracy of deep learning, a machine-learning technology, using ultra–wide-field fundus ophthalmoscopy for detecting rhegmatogenous retinal detachment , 2017, Scientific Reports.

[36]  K. Nishida,et al.  Risk factors for development of full-thickness macular holes after pars plana vitrectomy for myopic foveoschisis. , 2013, American journal of ophthalmology.

[37]  Hideharu Ohsugi,et al.  Accuracy of ultra-wide-field fundus ophthalmoscopy-assisted deep learning, a machine-learning technology, for detecting age-related macular degeneration , 2019, International Ophthalmology.

[38]  Hideharu Ohsugi,et al.  Accuracy of a deep convolutional neural network in detection of retinitis pigmentosa on ultrawide-field images , 2019, PeerJ.

[39]  Ning Qian,et al.  On the momentum term in gradient descent learning algorithms , 1999, Neural Networks.

[40]  P. Massin,et al.  PREOPERATIVE FACTORS INFLUENCING VISUAL RECOVERY AFTER VITRECTOMY FOR MYOPIC FOVEOSCHISIS , 2017, Retina.

[41]  Paul Mitchell,et al.  Prevalence and progression of myopic retinopathy in an older population. , 2002, Ophthalmology.

[42]  M. Treder,et al.  Automated detection of exudative age-related macular degeneration in spectral domain optical coherence tomography using deep learning , 2018, Graefe's Archive for Clinical and Experimental Ophthalmology.

[43]  Sergey Ioffe,et al.  Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[44]  Jitendra Malik,et al.  Analyzing the Performance of Multilayer Neural Networks for Object Recognition , 2014, ECCV.